We are nearing the last phase of our plan to address Panda, the Google search algorithm change that has hit HubPages. I’ve shared a number of major changes we’ve made in prior blog posts, but, to reiterate, we have been working on a plan to lift the punitive aspect of Panda from HubPages, where search engine rankings are being negatively impacted regardless of the quality of the Hub. Open publishing platforms like HubPages, by virtue of their openness, attract a wide range of content quality. As part of the Panda update, the presence of low quality Hubs has the ability to bring down rankings for all authors.
As we work to remedy this and restore traffic to high-quality Hubs, we take all changes to HubPages very seriously and have great respect for HubPages’ authors. Many of you have expressed feedback about Panda, about the disruption and the changes we are making, and about how we are going about it – both good and bad. Our sole focus is to create the most rewarding place for people to publish online that focuses on earning an income by providing information and interesting content to those that seek it. We will emerge from this as a stronger community and thank all of you for the feedback and thoughts. Part of that feedback has been to be given a heads-up on the changes we plan to make, so keep reading for a recap and our next steps.
Our plan has focused on four things to improve traffic to high-quality authors. First, we have improved quality on HubPages across the board by only allowing original content, we reduced the ratios of syndicated content to original, tripled the size of the moderating team to remove substandard content, and improved the type of sites we allow people to link to and promote. Even if weren’t for Panda, we believe these steps are an overall positive to the future for people that choose to participate on HubPages as they improve the overall experience with the site. Second, we have made architectural and other changes under the covers of HubPages to get rid of pages that may be construed as thin. These include noindexing pages like tag pages, forum posts and questions without responses, and implementing the rel=”author” tag. Third, we’ve spoken out against Panda’s effects on HubPages’ open platform compared to Google’s own properties, and we’ve reached out to Google in the official forums, and to Google employees directly. This effort has been fruitful in getting feedback on our overall plan. Our changes line up with the feedback we’ve been given from Google directly and their communication to us should allow us to better serve all of you.
We are now going to start testing moving individual authors to subdomains (e.g. pauledmondson.hubpages.com). This is a major development effort that is underway. Putting authors on a subdomain clearly delineates between sets of Hubs by author, so one author’s Hubs won’t negatively impact another. Paul Deeds, Jay Reitz and Tim Martin are working as quickly as possible to prepare the test. We are shooting for the last week of June to start the test on the site. We have the infrastructure in place to evaluate the results and will then consider a broader roll-out. We expect Hubs to get reindexed on the subdomains over a period of several weeks. Ideally, high-quality Hubs will see an increase in rankings and traffic. We have evidence on a small scale that this is the case and that it warrants further testing and a site-wide rollout if results are promising.
We are optimistic that the previous changes coupled with subdomains will allow high quality authors to prosper. We will report back on the changes as soon as we can evaluate them. We may be looking for a few participants to volunteer their accounts for testing. If you’re interested, leave a reply in the comments below (be sure to include your username).
83 replies on “Preparing a Test to Further Address Panda”
Despite the good, bad and the ugly of Panda, it’s clear HubPages has been doing what it can behind the scenes to get us all through it. I know we haven’t all gotten the answers we’ve wanted from Google, from the HP staff and from revising hubs, but if I have learned anything from my time here it is that we carry on. 🙂
I would love to be included in the testing of subdomains and how it impacts hubbers on my account carpesomediem. I look foward to seeing how successful this change will be.
If you find the hubs of Wilderness to be of the quality and type that you are looking for I would not be adverse to volunteering for the test.
It sounds interesting and could be beneficial.
If you think that this would help the overall site and also increase income, I will volunteer. Will just selected hubs be introduced or is it an all or nothing proposition. I am still busy editing older hubs to bring them up to the new standards.
As long as you have redirects in place for existing links, I volunteer,
Early Post Panda i chose to make a change to subdomains for my own general topic projects. Changing by author rather than topic is an interesting choice, Im game to guinea pig.
It’s the entire account that will need to be updated for the test.
In this situation would all of an author’s existing hubs be moved to the subdomain or could he/she keep the hubs on HP main site and begin the experiment by writing new hubs on the subdomain?
Thanks – I believe you answered my question while I was writing it 🙂
I renege on that volunteer. Sub domains should be by topical relationship not author. Especially for any possible type in traffic to an article – author names werent selected with the intention of being subdomains nor were most author accounts designed to stand alone.
Will this ruin all of the links that already exists to our current hubs?
I volunteer for the experiment.
Feel free to include me in your test of subdomains, if you believe that can be helpful.
Sunforged, the subdomain name can be different from your username (but it will be by author, not topic).
I don’t know if any of my stuff would be of the type of quality you’re looking for; but as far as I’m concerned, I’d be happy to be part of that kind of test.
Fabulous. Thank you!
I really don’t understand how this new step is a positive one. Why separate all of our articles from the rest of the site? What benefit is the hubpages domain then?
@sunforged, we are still working through how subdomain names will be selected. I just used my name as an example since I tend to select my name as a subdomain.
We considered subdomains by topic and authors, and folders by both as well, but the feedback from Google was subdomains by author.
Yes, the plan is to 301, but also to test new accounts. I saw your post in the forums on subdomains – it’s important to remember this is a test expanding on a few things we’ve tested already. This however is going to be larger.
It will help , if you would like me to respond to put comments here or in the official forum thread here http://hubpages.com/forum/topic/77245
Also, new accounts will be part of the test.
Wow! Impressive post Paul. I would be happy to volunteer with whatever it is HP needs during the testing trials. I have remained faithful that the beginning of a HP recovery would come, I am happy to see you sharing with the community the newest schematic for just this. 🙂
I wish you and all of the HubPages staff the best of luck.
Respectfully in solidarity~
K9keystrokes on HubPages
My husband also has an account. He has been a hubber for over 3 years. His hubs ( numbering just above twice my number of hubs) get 4 times the traffic. We feel the quality is the same and in the same general topic.
It would be interesting, and hopefully profitable, if your subdomain changes could test my hubs to see if it pulls me back closer to his in relative traffic terms. Of course he would not mind his being tested as well.
Will authors be able to change their user hame given that a keyword included in a domain is an important seo factor?
Clearly from Sunforged’s comments, this will only be effective if the sub-domains are based on topics, not on authors, therefore those of us who write on multiple topics, e.g. pets, gardening, cooking etc cannot expect this to enhance our readership or rating within Google!
As I said in the forums, I’m available for testing.
Its good to hear that the subdomain concept isnt locked to pre-existing author names. Its a big move. There will be a land rush for ‘premium” keyword based subdomains – fitting these general topic author accounts into custom subdomains isnt going to be easy either. May I reserve Mesothelioma.hubpages.com if I volunteer, lol. I kid, but thats gonna be a whole new escapade!
I can understand originally designing a site structure or new accounts to this model, but has any large site ever done such a massive “tweak” of site structure?
Hi, I have two questions. Firstly, I have always been told that Google hates subdomains. Isn’t that why websites pay to have domain names instead of choosing free subdomain hosting?
Secondly, are all accounts by the same author going to be combined into the same subdomain?
I’m sorry to ask, I’m sure the answers to both questions are obvious to those who have been at this longer than I have.
I think the idea of subdomains is likely to be a very positive one. I can only guess that because of some of the dross/spam that has been on HP previously, this has affected all our hubs. My feeling is that providing we, as individuals, don’t post dross, we will be much less affected by other hubs.
Time will tell but it does sound like a positive move and given my low hits I’m happy to volunteer too.
Very good info here and I appreciate all of hubpages efforts regarding Panda. I’m not in total agreement regarding all the changes here but Google is getting what they want which was to try and scare sites into self monitoring their content.
Panda didn’t effect the rankings of any of my hubs and I’ve been lucky enough to have most of my hubs rank 1, 2 or 3 for very high volume keywords and as a result earned thousands of dollars for which I really only have Hubpages to thank.
The problem for me is that after Panda my Hubs still ranked at the top and Google gave them more love than ever. But, hubpages has unpublished quite a few regardless of any changes I’ve made. I have pretty much given up trying to figure out the plan here and that’s ok. But I would really like to point out that some Hubs were adored by Google and rewarded after Panda and we may be shooting ourselves in the foot and over reacting just a tad.
Any thoughts would be appreciated. I wrote most of my Hubs thinking an update like Panda was inevitable and prepared for it. I didn’t lose rankings but actually gained and while Google rewarded me I pretty much got “hub slapped”.
I’m also available for testing because I’ve had some very good luck in the Search Engines and may be able to help find solutions to help us overcome this problem.
Whilst I welcome any attempt to regain traffic, I am very wary of this approach. It effectively means that we will all have blogs within Hubpages, I know very few people who manage to successfully monetize on Blogger, for example.
I am much more attracted to the idea that Sunforged has, which is to separate by topic. For me there is no better way to determine precisely what topics are being frowned upon by Google, and what topics are deemed acceptable.
I can’t think of any other multi-author site which uses the author subdomain model, is there any precedent set here? A positive case study?
About.com seperates topics by subdomain, and their quantcast figures, whilst estimated…. appear to show that the effect of Panda on their site has been pretty minimal compared with Hubpages. 46m uniques per month, down to 40m uniques per month. What’s that? 15%? Their response to a decline in traffic was to increase the number of subdomains in popular categories, they had 800 pre-panda, and were going to increase that by 25%.
That would, for me, seem like the most logical way to restructure. It would then be possible, in the event of future updates, to identify affected niches and unaffected niches. If Panda 5.8 in 2015 results in a 10% reduction in traffic, and webdesign.hubpages.com gains 5%, whilst diet.hubpages.com and photography.hubpages.com were each to lose 30%, you would know precisely where to concentrate your moderation efforts. Furthermore, this would enable the possibility of bespoke content regulations on a topic by topic basis.
Please don’t take this as being critical, the past is the past, I only concentrate on the future. Such a system would have enabled you to analyse data much more effectively following Panda 1 and Panda 2.
The question that I should perhaps be asking is, why not trial both methods at the same time? Why not create subdomains for say 10-20 topics, of low importance to Hubpages and hubbers revenue wise, at the same time as trialing the subdomain by author model? Would this be difficult to undertake?
I too would prefer to be left out of this experiment if possible, at least until the stage that you are confident that it improves performance.
I’m willing to volunteer as well – Lily Rose
My username is : nanospeck . I would love to be a part of this change and in necessary case am ready to offer any kind of programming assistance that could help the entire process.
Count me in.
I don’t know, if our profiles are going to be in subdomains, we are going to be treated by search engines as independent websites rather than a part of hubpages. I’m not sure if that’s good or not. I thought we are quite visible in search engines because of ‘hubpages’. If we are going to be treated as an independent site, I’m not sure if search engines will still pick us and our hubs.
Please leave me out of testing phase until you are sure.
I am happy to be involved in the sub domain testing, my account name is “psf”
Question. Would our pages still show up in Hubpages like they do now or would they be separated and stand alone like a separate blog? Would hate to lose contact with many of the HubPage writers that I have gotten to know through this venue.
@Ryan Kett. We did consider subdomains by topic and originally proposed that idea. We talked with many sites that are structured this way and all the data we got back suggested that after a site got hit by panda, it brought traffic down in all subdomains when organized by topic. About.com may be the best example of a site with the least impact at scale. About’s content is created by many less contributors and its quality is less varied as well. We believe these are pieces of Panda that would still negatively impact HP if organized by topic. The examples we have of sites that are broken down at a more granular level are the ones that held or improved.
It’s important to remember we are starting with a test. We will share data and the larger plan on how subdomains will be allocated in the coming weeks.
Really good thoughts and appreciate the feedback.
I will be glad to volunteer, if my hubs are suitable.
thank you for all the effort put in to helping us all out, it is much appreciated!
I came back to mention the prospects of a gold rush for subdomains. That does need some serious consideration when / if you decide to implement this.
Otherwise we can expect to see people setting up hundreds of accounts with 1 hubpage on each. bestgreenwidgets.hubpages.com, bestredwidgets.hubpages.com, bestpinkwidgets.hubpages.com.
That doesn’t bode well for the hubber who is getting 50 views a day to a Hubpage entitled The 5 Best Red Widgets, and I wouldn’t want to have to register 30 subdomains to protect my most valuable Hubs from those who steal the title to create their subdomain, particularly when author pages seem to establish authority so easily.
I should really try to get these things down on one post, all in one go, but I have thought of something else that you need to consider.
And that is trademark breaches. Whilst you are permitted to use a trademark in an article title, and in a URL structure as a file name, there is precedent (lot’s of it) for people being sued for using a trademark in a second level domain. Google were sued for it I believed, Apple clearly state that they would sue for this breach.
You could not, for example, have ipod.hubpages.com, or google.hubpages.com.
So Hubpages would need to check every single username against the USPTO database before converting it into a subdomain. Not sure if they have some sort of API which enables that to be done by way of an automated process, but I doubt it.
In the cases of a username containing a patented name, you would need the user to change their username, or to give them a subdomain which is unrelated to their chosen username.
For example, this hubber is called Super Chef: http://hubpages.com/profile/Super+Chef
A search for “Super Chef” on TESS found numerous live trademark registrations for that name. That means that you could be liable should you use the subdomain superchef.hubpages.com
That, for me, is the most challenging aspect of such a change.
I’d be curious to take part in the described test. My account name is listed to the right.
I am interested in the points that Ryan and sunforged have made with regards to the subdomains. It does seem that using the nomenclature of their username (or their choice of subdomain if not) may cause some problems. However, using subdomains of categories would seem to almost defeat the purpose. If you are trying to weed out bad content from bad hubbers then that content will still be present in the category subdomains (unless it is weeded out beforehand). So it seems to be a tricky situation.
In any case, after consideration I would like to be included in the testing.
..In the example, super-chef is the third level domain.
SORRY, my mistake, please disregard that. I simply googled “subdomain trademark” and didn’t read the results properly, my bad. It looks as if there hasn’t been a case since somebody tried to sue Google, which was settled out of court, and as such there is no precedent and this clearly isn’t an issue.
Thanks for answering my post Paul, the responses which followed crossed with your response whilst awaiting moderation.
Cheers and good luck with the trial.
If you still need more volunteers, I’ll be glad to participate with my Sherri L Souzen profile.
Thanks so much for all you are doing to save the site. I really enjoyed working here, and though I took a break, I will be returning to write at least one article a week in the future. I hope to contribute something good to the all new, and improved, Hubpages domain.
I will be happy to be a volunteer though my hubs may not fulfill the quality requirements.
I would love to be part of this team!
Just saw this post and have read all of the comments so far.
If you are still looking for volunteers I too would be willing to volunteer.
I have a large number of hubs on a wide range of topics but, after reading all of the above comments to date think that the username is probably the best choice for this. My username is Chuck.
I would love to be included for the test. Count me in 🙂
The Testing sounds like a good way to go. I don’t want to volunteer however, but I am reading this set of comments with great interest. Brilliant updates here too! The upgrades and improvements lately have been going in the right direction. There are big differences between About and Hub Pages, and the differences are enough to warrant trying it your way, Paul. I hope it works, and Best of luck to everyone! And excellent feedback here too! Cheers!
Some really interesting responses here. Please, check if there is any misuse of usernames before doing anything, especially among those volunteering.
As for moving stuff to subdomains, definitely leave mine out of it. Thanks.
I will be a guinea pig! 🙂
Count me in, if it’s not too late….
I am ready to volunteer my account for testing subdomains. I have strong belief that this will work out very well. Our daily visits will improve back to post panda period. Thanks.
My account like the rest is available for you to test with. I am however a writer who writes more for personal reasons to get out a message rather than make a grand profit.